The Knowledge Factory Framework is a novel ML-centric platform to capture information and build knowledge across the entire data landscape. The system leverages challenges such as resume mechanism from a model failure, restart mechanism for unprocessed files backed by plug and unplug mechanism for new models’ addition or existing models’ updates, knowledge repository curation, adaptability, scalability, and most importantly the mechanism for reusing the existing knowledge to enable the extraction of more information from the data.

Peptides represent a growing therapeutic space between small molecules and biologics. More than 80 peptide drugs have reached the market for a wide range of diseases. Due to their size and special composition, peptides can be approached from an AI perspective in a different way than small molecule drugs. In this presentation, we show the newest predictive and generative models for peptide discovery.

Expert Systems has developed DrugInteLLM, a suite of LLM experts that are specifically designed to address different tasks and activities related to early drug discovery. PharosGPT (targets, diseases, and ligands) identifies target-disease associations. With litGPT, we process subsets of the literature and patent corpus. ChEMBLGPT queries databases for bioactive compounds. Our machine learning platform has thousands of target-based models and hundreds of ADMET-based models, which are interfaced to ActivityGPT (which is tasked to predict bioactivity endpoints). We will briefly describe our platform for target and ligand selection, and how GPTs can support drug discovery.

The advent of artificial intelligence (AI) and cloud computing presents an opportunity to innovate drug discovery by fully leveraging data-driven modeling to reduce the number of cycles necessary to yield a clinical candidate. Here we will present the Predictive Insight Platform (PIP), a cloud-native modeling platform developed at AstraZeneca with a special emphasis on its architecture, integration, usage, and impact on drug discovery. We will also discuss challenges for creating a chemistry-aware autoscaling modeling platform and how we overcame those.

Recent years have seen increased talk around AI, both in industry and now consumers. Computer-aided drug design is a vital part of drug discovery, but we still see limitations; we cannot simply enter a target structure and get a hit drug. In this talk we’ll explore the practical strategies and tactics we propose are needed to deliver on the AI drug discovery vision.

In quantitative systems pharmacology (QSP), the identification of predictive features from intricate, high-dimensional multi-omics datasets is crucial for unraveling and addressing therapeutic responses. The integration of artificial intelligence (AI) methodologies has emerged as a promising solution to reduce the complex feature space in various data domains such as in vitro, in vivo, and clinical settings. Here we introduce a collection of innovative AI-driven tools specifically tailored for high-dimensionality reduction in QSP.

Artificial Intelligence has emerged as a transformative force in the field of drug design, promising enhanced efficiency, precision, and speed. In this presentation, we will delve into the challenges encountered in the drug design process and detail our approach to overcoming these hurdles by building and implementing new AI capabilities. We will showcase a selection of real-world projects, illustrating the impact of these techniques on the advancement of drug discovery.

As the industry continues its innovation spree across small and large molecules, Tech-enabled drug discovery vendors are re-establishing themselves as digital biotechnology companies with in-house therapeutic pipelines along-with platform-based and project-based partnerships to pharma players. Also, incorporating integrated AI-driven solutions in clinical trial design will aid in reducing costs, timelines and increasing efficiency through DCT models. The presentation will explore key trends and opportunities offered by AI/ML to different types of stakeholders across the drug discovery and development value chain.